Hello friend,
I'm going to make you a promise.
By the end of this little book, you'll know more about actually using Claude than 95% of the people out there paying for it. Not the theory. Not the hype. The real moves that take you from "I open it sometimes and ask it questions" to "this thing runs half my life now."
I'm not a developer. I have zero coding background. I figured all of this out the slow, confusing way... by clicking around, getting it wrong, and slowly piecing together what actually matters. This book is me handing you the map so you can skip the confusing part.
Here's what's actually at stake, and I promise I'm not being dramatic.
The exact same Claude is sitting on millions of screens right now. A tiny slice of people are getting unreal results from it. Everyone else is typing one-line questions and quietly wondering what the fuss is about. The gap between those two groups isn't intelligence, and it isn't coding. It's a handful of setup moves nobody tells beginners about. And a year from now, the people who made those moves early are going to look like they have superpowers. I really believe that.
This book is those moves. So before we start, let me tell you what's coming, just so you keep reading:
- →There's one 20-minute setup in Chapter 2 that I'd trade every other trick in this book to keep. It's the thing that makes Claude stop treating you like a stranger every time you open it.
- →There's a single habit in Chapter 4 that's saved me more hours than anything else I do with AI. Most people never learn it, and it's almost embarrassingly simple.
- →And there's one prompt, coming up soon, that gets Claude to build your setup for you, by interviewing you. You don't even have to know what you want yet. It figures it out.
Oh, and one thing you're almost certainly getting wrong already: there are three different "Claudes," and most beginners spend all their time in the one they should barely be using. We fix that in about two minutes.
Stick with me. Here's how the book works:
- •5 short chapters. Each one is a single skill that builds on the last.
- •Explained like you're in 3rd grade. Not because you're dumb. Because clear beats clever, every single time. No jargon without a translation.
- •Copy-paste prompts in every chapter. Little boxes you can hand straight to Claude. I tell you exactly when to use each one.
- •A "Go Deeper" box at the end of each chapter for when you're ready to level up. Skip them on your first read. Come back when you want more.
Read it once front to back to get the whole map. Then come back to each chapter as you're ready for it.
Let's get into it, shall we?
Start Here
First, the one idea that changes everything
Before chapter 1, you need the single mental shift that makes all of this click. If you only remember one thing from this whole book, make it this:
Stop thinking of Claude as a chatbot. Start thinking of it as the smartest new hire you've ever managed.
That new hire is brilliant. Genuinely. But on day one, they know nothing about you. Nothing about your goals, your job, your taste, your situation. They have no memory of yesterday. Hand them a vague task and they'll guess. Brief them properly and they'll outperform people you'd pay six figures.
That's the whole game. Every chapter in this book is really just one skill wearing different clothes: how to onboard your brilliant new worker so they actually know what they're doing.
Bad prompts? Untrained employee. No memory? Untrained employee. No tools, no saved systems? Same thing. Fix that, and "AI is kind of overhyped" turns into "wait... how did I ever work without this."
The Big Picture
Where this is all actually headed (the part that won't change)
I want to zoom out for one minute, because this is the most important thing in the book.
The last six months have been a firehose. So much AI advice, so many "you HAVE to try this" videos, a new must-use tool every week. A lot of it is genuinely good. But here's what I've realized watching all of it go by: the core concepts haven't changed the entire time. The tools change weekly. The foundation doesn't.
And the foundation is this. You're not learning to "use a chatbot." You're building an AI operating system around your life. A setup that knows you, works the way you work, and quietly does more and more for you over time. That's where the whole world is headed, and the people who set it up now are going to look like they have superpowers in a year.
It stands on three legs. The entire rest of this book is just these three, in order:
A context base so your AI knows you
One home it reads from every time, so it never starts as a stranger. Chapter 2.
Use AI to build AI
Let it interview you to uncover what's worth automating. It's better at spotting it than you are. Chapters 2 & 5.
Repeated work becomes a skill
Anything you do more than once, capture it once, and from then on it runs itself. Chapters 4 & 5.
That's the OS. Context so it knows you, AI that helps you find what to automate, and skills that make the automations permanent. Tools will keep changing. This is the part to actually understand, because it's the future and it's not going anywhere.
Okay. Zoom back in. Let's go build it.
Get Oriented
The 3 Claudes (and which one to actually use)
Quick thing that confuses every beginner: "Claude" isn't one thing. It's basically three, and people waste weeks not knowing which one they're supposed to be in.
Here's the whole picture in 20 seconds:
Chat
The normal chat window at claude.ai
Think of it as
Texting a brilliant friend
Cowork
An agent inside the desktop app that works in your files on its own
Think of it as
A new hire with their own desk
Claude Code
Claude with access to your whole computer, runs in the terminal
Think of it as
Hiring a whole dev team
Here's the part nobody tells you, and it's the most important sentence in this section:
For 99% of what you'll ever want to do, you should be living in Cowork.
Chat is great for quick questions. Claude Code is the deep end (we'll talk about it at the very end as your "graduation"). But Cowork is the sweet spot, and it's getting scary good. It's an actual employee that reads your files, does multi-step work, and hands you finished stuff... all inside a friendly app, no terminal, no code. That's where this whole book lives.
So here's your move before chapter 1: download the Claude desktop app. Mac or Windows. That's home base.
Alright. You've got the mindset and you've got the app. Let's get you into your first real session.
Chapter 1
Your first Claude session
The big idea: The quality of what Claude gives you is almost entirely decided by the quality of what you give Claude. Learn to brief it like a pro and everything else in this book gets easier.
Most people type into Claude like it's a Google search. "Write me an email." "Help with my marketing." Then they get a generic, mushy answer and think the tool is overrated.
It's not the tool. It's the brief. You gave a vague instruction, so you got a vague result. Garbage in, garbage out isn't a Claude problem. It's a briefing problem. And it's the easiest thing in the world to fix.
The only prompting formula you need: R-C-C-F
Every strong instruction has four parts. Role, Context, Command, Format. Once you see it, you can't unsee it.
- •Role — who you want Claude to think like. "You're a direct-response copywriter who's written for huge brands." This sets the whole frame: the vocabulary, the priorities, what "good" even means to it.
- •Context — the actual details. Who's it for, what's the goal, what's the background, what should it avoid. This is where 90% of the quality hides, and it's exactly the part everyone skips.
- •Command — crystal clear on the one thing you want. "Write the email." Not "write the email, design the funnel, and pick a price." One job at a time.
- •Format — what should the output actually look like? Length, structure, tone, bullets or no bullets. Don't tell it the shape and you'll get Claude's default shape, which is rarely the one you wanted.
Watch what that does.
Before
write me a cold email for my consulting service
After
Role: You're a B2B sales writer who specializes in short, high-reply cold emails.
Context: I run AI consulting for accounting firms. My prospect is a managing partner at a 30-person firm, slammed during tax season, skeptical of "AI hype." My angle is saving their staff 10 hours a week on document review.
Command: Write one cold email that earns a reply, not a sale.
Format: Under 90 words. No fluffy greeting. One specific claim, one soft question to close. Plain text.
Same Claude. Completely different result. The second one took 40 extra seconds to write and saved an hour of editing. That's the trade every single time.
Your first session, step by step
- Open the Claude desktop app. Make sure you're in Chat mode for now (top of the window).
- Pick something real and small you actually need done today. A reply, a plan, a rewrite. Don't practice on a fake task.
- Write your request using R-C-C-F. Don't overthink it. Even getting Context right is 80% of the win.
- Read the answer like a critic, not a customer. Too long? Wrong tone? Just tell it. "Cut it in half." "More casual." It remembers the whole conversation, so you're sculpting, not starting over.
That back-and-forth is the skill. You're not looking for a perfect first answer. You're steering toward a great final one.
Copy-paste prompt — The R-C-C-F upgrader
Use this when you've got a lazy one-liner and want Claude to turn it into a real brief before answering.
I'm going to give you a rough request. Before you answer it, rewrite it for me using the R-C-C-F format: Role, Context, Command, Format. Ask me any questions you need to fill in missing Context. Once it's a strong brief, THEN do the task. Here's my rough request: [paste your messy one-liner here]
The beginner mistake to avoid
Two extremes, and people fall into both. The first is a wall of context with no clear command, so Claude doesn't know what you actually want. The second is a sharp command with zero context, so it confidently gives you a polished, wrong answer. You need all four parts. Missing context gets you confident-and-wrong. Missing format gets you right-but-in-a-shape-you-can't-use.
Go Deeper
Stop typing your prompts. Start talking them. The real reason people give thin context isn't laziness, it's friction. Typing a full brief is annoying, so you cut it short. Install a voice-to-text tool (I use Wispr Flow) and just talk through the whole situation the way you'd brief a coworker. Your inputs get 3x richer automatically, because talking is easier than typing. This one change quietly upgraded everything for me.
And the prompt that makes Claude prompt itself:
Copy-paste prompt — Make Claude prompt itself
You're a world-class prompt engineer. I want to accomplish [your goal]. Interview me with a few sharp questions, then write me the ideal prompt I should be using to get the best possible result. Show me the final prompt in a copy-paste block.
Chapter 2
Build your AI second brain
The big idea: In Chapter 1 you learned to talk to Claude well. The problem is it forgets you the second you close the chat. This chapter gives Claude a permanent memory of who you are, so it shows up to every task already knowing you. This is the one that feels like magic.
Here's the thing that quietly wrecks most people's experience: Claude is a stranger every time you open a new chat. You explain who you are, you explain your work, you get a great answer. Tomorrow? Fresh chat, blank slate, explain it all again. That's not using AI. That's re-onboarding the same employee every single morning.
We're going to fix that for good. And we're doing it in Cowork (remember, the agent inside the desktop app, where you should be living).
A "second brain" sounds fancy. It's not. It's just a folder on your computer full of notes about you that Claude reads before it does anything. Who you are, how you work, how you write, what you care about. Once it exists, every task starts with Claude already briefed.
We build it in two moves.
Move 1 — Give Claude standing instructions
This is the step 99% of people skip. Cowork has a Global instructions box that Claude reads at the start of every single task, forever. It's the standing brief you'd give a new hire on their first morning.
- Open the Claude desktop app. Switch the top mode from Chat to Cowork.
- Go to Settings → Cowork → Global instructions → Edit.
- Delete whatever's in the box.
- Paste the prompt below and fill in the
[brackets]so it's actually about you. - Click Save.
Copy-paste prompt — Cowork standing instructions
Use this once, right now. It's the standing brief Claude reads before every task, forever.
# Who I am My name is [your name]. I'm a [your role] at [your company / "myself"]. I use Cowork to [the main thing you want, e.g. "draft content, run my admin, and produce client work"]. # My folder system I keep one working folder for Cowork. Inside it: - ABOUT ME: context files about me, my work, and my voice. - OUTPUTS: finished work. - TEMPLATES: reusable structures for stuff I make often. Rules: - At the START of every task, read every file in ABOUT ME before doing anything else. That's how you know who I am. - Don't read OUTPUTS or TEMPLATES unless I point you at a file. - Save every finished deliverable into OUTPUTS, in a subfolder named after the project. # How I work - Tone: [how you want it to sound, e.g. "direct, plain, no corporate jargon"]. - Lead with the answer, then the detail. Short sentences over long ones. - When you're unsure what I want, ASK one clear question instead of guessing. # Rules you don't break - Never delete, overwrite, or move a file without asking me first. - Never invent a fact, a number, or a source. If you don't know, say so.
Move 2 — Let Claude interview you to build the brain
Now the magic part. You're going to make a folder, then paste one prompt that turns Claude into a sharp onboarding interviewer. It drills you with questions about your work, your voice, your tools... then writes all of it into files inside that folder. That folder is your second brain.
- Make a new folder on your desktop. Call it
Claude Cowork. - Drag it into the Cowork window. When it asks for access, click Allow.
- Paste the prompt below and hit enter.
- Answer its questions in detail. This is the whole game. Lazy answers get you a lazy assistant. Specific answers get you one that sounds like you and knows your business.
Copy-paste prompt — The second-brain builder
Use this right after Move 1, in your new Claude Cowork folder. Answer everything it asks.
You're going to help me build the "operating system" for working with you in this folder. The goal: context files you read at the start of every task so you always know who I am, how I work, how I write, and how I want things done. Work in two phases. Do not skip ahead. ## PHASE 1: Interview me Ask me 2-3 questions at a time, not a wall. Wait for my answers. Go DEEP: when an answer is vague, ask a follow-up until it's specific. Use plain language. Cover, roughly in order: 1. IDENTITY: my name, role, business, what I actually do all day. 2. RESPONSIBILITIES: the main things I'm on the hook for. 3. GOALS: what I'm trying to make happen this quarter and year. 4. THE REPETITIVE WORK: tasks I do over and over that I'd hand off first. 5. VOICE & WRITING: how I want things to sound. Ask me to PASTE 1-3 samples of writing I'm proud of, then tell me the patterns you notice and confirm them. 6. TOOLS: the apps I work in every day. 7. PEOPLE: who I work with and how I refer to them. 8. PET PEEVES: phrases, formats, and behaviors I never want from you. Don't let me off the hook on voice or the repetitive work. Those matter most. ## PHASE 2: Build the files When you have enough, build this exact structure inside this folder: Claude Cowork/ ├── CLAUDE.md (master briefing, under two screens) ├── ABOUT ME/ │ ├── aboutme.md (identity, responsibilities, goals, people, tools) │ ├── brand-voice.md (exactly how I write, with the patterns from my samples) │ └── working-preferences.md (how to work with me, my pet peeves, hard rules) ├── OUTPUTS/ (empty for now) └── TEMPLATES/ (empty for now) Everything must match my voice and rules. When done, summarize what's in each file and tell me to confirm my global instructions are saved.
When it finishes, open the folder. You'll have a real second brain Claude reads on every task. From here on, it shows up already knowing you.
The beginner mistake to avoid
Rushing the interview. The folder is only as smart as the answers you give it. If you mumble through the voice and repetitive-work questions, you get a generic assistant. Spend the 20 minutes. Paste real writing samples. This is the highest-leverage 20 minutes in the whole book, and you only do it once.
Go Deeper
Your second brain isn't frozen. Anytime Claude does something the way you love (or the way you hate), just say "add that to my working-preferences file so you remember it next time." It updates the brain on the spot. Over a few weeks it gets eerily good, because it's learning your taste in real time. The setup is the start. The compounding is the point.
Chapter 3
The MCP servers + plugins worth installing
The big idea: Up to now, Claude only knows what's in the chat. But your real work lives everywhere else: your email, your bank, your meetings, your design tools, the live web. This chapter gives Claude hands so it can reach into all of it and actually do things, not just talk about them.
Quick translation, because the name is scary and the thing is simple. MCP stands for Model Context Protocol. Forget the words. Here's all it means:
MCP is the plumbing that connects Claude to other tools. In plain English, it gives Claude hands.
Without it, Claude is a brilliant brain in a jar. With it, that brain can read your email, search the live web, make images, pull your bank transactions, read your meeting transcripts, and more. Same Claude. Suddenly it can reach the rest of your life.
The one rule that matters
Here's the whole philosophy, and it's simpler than any list I could give you:
If you use a tool every week, connect it.
That's it. The goal is for Claude to reach into your actual work life, not a curated demo version of it. Every tool you plug in is one more thing Claude can do for you instead of you doing it by hand. So the real answer to "which ones?" is "the ones you already use." Below is my personal stack to show you what that looks like in practice. Yours will be different. Copy the pattern, not the exact list.
My actual MCP + connector stack
Claude reads, drafts, and acts on your real email, calendar, and Drive
Searches today's internet, with sources. Claude's training has a cutoff; this fixes it
Crawls any site so Claude can study competitors and pull real data off pages
Want images? This is the one. Thumbnails, product shots, ad creative, the whole thing
Records your calls and gives Claude every past transcript to pull from
Claude draws out flows, diagrams, and visuals for you
Build node-based, multi-step workflows when you're ready to automate bigger things
Claude can read transactions and help with the finance side
Claude reaches into your team conversations
Claude reads and writes to your knowledge base
Claude works with your real pipeline and contacts
Don't install all of these on day one. Start with the two or three you'd miss most (for most people that's email + Perplexity + one tool your job runs on), then add the rest as you go. The list isn't the assignment. Connecting what you actually use is the assignment.
How to actually add them
Two flavors, both easy:
- •One-click connectors: in Cowork, hit the + near the message box, open connectors, and sign in. Many of the big ones (Google, Slack, Notion) are right there.
- •Custom MCP servers: for the rest, go to Settings → Connectors → Add custom connector. Each takes about two minutes. And if the setup ever looks intimidating, you don't have to figure it out alone, use the prompt below.
Copy-paste prompt — Let Claude install it for you
Use this when you've found a tool you want to connect but the setup looks technical. Claude walks you through it like a patient teacher.
I want to connect [tool name] to you so you can [what you want it to do]. Walk me through the exact setup step by step for the Claude desktop app, one step at a time, and wait for me to confirm each step before the next. Assume I'm a complete beginner and explain anything technical in plain English.
The beginner mistake to avoid
Connecting everything at once and clicking "allow" on every permission you see. More access isn't more power. It's more risk for zero benefit until you actually use it. Add tools as you reach for them, and glance at what each one is allowed to touch. Start with two or three. Grow from there.
Go Deeper
The real magic shows up when you chain tools in one breath. Try: "Use Firecrawl to pull the latest from [competitor], use Perplexity to find what's trending in our space this week, then use Higgsfield to make me three thumbnail concepts based on what's working." That's three tools, one sentence, a full mini-workflow. That's the moment Claude stops being a website you visit and becomes the command center your whole work flows through.
Chapter 4
How to build your own Claude skill
The big idea: A skill is how you teach Claude to do something your way, perfectly, every time, with one command. It's the move that turns Claude from a tool you use into a system that works for you.
Here's the honest 10-second version of what a skill is, before anyone makes it complicated:
A skill is a saved set of instructions Claude follows on demand. You do a task well once, save how you did it as a skill, and from then on you type one command and Claude runs the whole thing the exact same way. No re-explaining. No drift. Your good process, frozen and reusable.
Technically it's just one folder with a file called SKILL.md inside. That file says: "when this kind of task comes up, here's exactly how to handle it." No code. Just instructions Claude reads and follows.
Think about what that compounds into. The competitor research you figured out last week. The way you like your emails written. Your exact invoice format. Each one becomes a one-word command instead of a 20-minute re-explanation.
The trigger that builds your skill library on autopilot
You don't sit down and "write skills." You let them pile up from real work. Here's the rule, and it's one of the three legs of the whole AI operating system:
Anytime you do something more than once, and you know you'll do it again, make it a skill.
That's the entire trigger. Not "someday I'll automate this." The second time you catch yourself doing the same task, you stop and capture it. Honestly, the skill-creator skill and this one habit have saved me so many hours it's not even funny. Work I used to redo every week now runs itself with one command.
So anytime Claude does something useful that you'll want again, say this:
Copy-paste prompt — Turn anything into a skill
Use this the moment Claude nails a task you know you'll repeat. This is the single most valuable habit in the book.
That worked great. Write this up as a skill file we can save, so we can repeat this exact process every time with one command. Give it a clear name and a clear "use this when..." description so you know when to reach for it later.
That's it. Do the task once, fire that prompt, and you've got a reusable skill. Do that for a month and you'll have a personal library that makes Claude feel custom-built for your life. Because it is.
Where skills live (so you can actually install them)
Same file works in both places:
- •Claude desktop / Cowork: Settings → Capabilities → Skills → Add skill, then upload the folder (or a zip of it). (Skills need "Code execution and file creation" turned on in Settings → Capabilities first.)
- •Claude Code: drop the skill folder into
.claude/skills/in your project. Done.
The beginner mistake to avoid
Building one giant skill that tries to do everything. Skills work best small and focused, with a clear name and a clear "use this when..." line. That's how Claude knows when to reach for it. Five sharp skills beat one bloated one.
Go Deeper
Install Anthropic's official skill-creator skill (it's free, in their skill library). It's a skill whose entire job is helping you build other skills. You describe what you want to automate in plain English, and it writes the properly-structured skill file for you the first time. It's the fastest way to go from "I do this a lot" to "Claude does this for me now." Use it to build your second skill, and your third, and your tenth.
Chapter 5
The exact skills to install first
The big idea: There are a lot of useful skills out there. But the ones that'll actually change your life depend entirely on what you do all day. So the smartest first move isn't copying someone's list. It's letting Claude interview you to uncover your highest-leverage skills, then installing a few that help almost everyone.
Here's the trap I want you to skip. You'll see big lists of "the 35 skills I use" (I have one too) and feel like you're supposed to install all of them. Don't. Most of those are specific to my work. The skill that saves me two hours a day might be useless to you, and the one that'd save you two hours isn't on anyone's list yet, because it's about your job, not mine.
So we start the right way: we ask.
Let Claude interview you to find your skills
The best skills are personal. The fastest way to find yours is to let Claude look at how you actually work and tell you where it can help. Paste this and answer honestly.
Copy-paste prompt — Uncover your most useful skills
Use this first. It finds the skills that are actually worth building for your specific life and work.
I want to figure out which Claude skills would save me the most time and headache. Interview me to find them. Ask me 2-3 questions at a time, not a wall. Dig into: - The tasks I do over and over every week (be specific, walk me through them) - The work I dread or procrastinate on - The stuff that eats my time but doesn't really need my brain - The things I wish would "just happen" without me When you understand my work well enough, give me a ranked list of the 5-7 skills I should build or install first, easiest-and-highest-impact at the top. For each one: what it does, why it fits me, and whether it's something we build with skill-creator or something I can install. Then offer to build the top one with me right now.
That list will be worth more than any generic one, because it's yours. Then use Chapter 4's habit ("turn this into a skill") to build them one by one.
The few that help almost everyone
While your personal list will be unique, a handful are close to universal. I'd install these no matter what you do:
1. skill-creator (Anthropic official, free)
The skill that builds other skills. Install this first, full stop. It's how you'll build every skill the interview uncovers. Describe what you want to automate in plain English, and it writes the properly-structured skill for you.
2. A Gmail triage skill
Non-negotiable in my book. It scans your inbox and sorts everything into respond, delete, ignore, then clears the junk only after you approve. If you get more than a handful of emails a day, this hands you back real time every single morning.
3. last30days (free, open-source, 36k+ GitHub stars)
The single most useful skill I've found for content and research. You give it any topic and it searches all of social media at once (Reddit, X, YouTube transcripts, TikTok, Hacker News, even Polymarket odds) and hands back what real people are actually saying in the last 30 days, ranked by traction. No tool you're already paying for can do this. If you make content, do research, or just want to know what's really happening on a topic, this is gold.
The beginner mistake to avoid
Installing 30 skills on day one because a list told you to. You'll never use 25 of them and you'll forget what the other five even do. Run the interview, install the three above, actually use them for a week, then add skills only when a real, repeated task asks for one. A skill you use beats ten skills you don't.
Go Deeper — three I've already built for you
If you want plug-and-play wins while you figure out your own stack, I packaged three beginner skills you can install in two minutes:
- •Memory — Claude saves what it learns about you and reads it back every new chat, so it stops forgetting you. (Week one it knows your name. Month three it knows your whole operation.)
- •Reorganize — point it at a messy folder and it sorts, renames, and files everything, asking before it moves anything.
- •Pause + Resume — say "pause" when you stop and "resume" when you're back, and Claude picks up exactly where you left off.
Memory makes Claude know you, Reorganize makes it know your stuff, Pause+Resume keeps your momentum. Grab them, then let the interview build the rest.
Where To Go Next
Graduating to Claude Code
You did it. And look at what you actually built. It wasn't five random tricks. It was the three legs of your AI operating system: a context base so Claude knows you (Chapter 2), AI that interviews you to find what's worth automating (Chapters 2 and 5), and skills that make those automations permanent (Chapters 4 and 5). Plus the hands to reach your whole life (Chapter 3) and the mindset to run it all (Chapter 1).
That's the thing that won't change while the tools keep churning. You're not someone who "uses AI" now. You've got a system that compounds. That genuinely puts you ahead of almost everyone using this tool.
Here's the honest truth about where to go from here: for about 99% of what you'll ever want to do, Cowork is enough. And it's getting better every month. Don't let anyone make you feel like you need to touch a terminal to be "real" at this. You don't. Live in Cowork. Get great at it.
But there's one more door, for when you're ready.
Claude Code is the deep end, and it's where the ceiling disappears entirely. It gives Claude access to your actual computer: it can build software, automate real workflows, and complete genuinely autonomous tasks. I built things with it, with zero coding background, that I'd have had to hire a developer for. A real example: I asked it to look through my downloads folder and delete anything older than 30 days. A few minutes later, done. That's the small version. The big version is custom tools and entire workflows you describe in plain English and it just... builds.
And here's the beautiful part: everything in this book pays off there at full strength. Your R-C-C-F habit makes your instructions land. Your second brain gives it context on every task. Your skills run your systems with one command. You're not starting over. You're leveling up.
When you're ready for that door, I've got you. But for now? Go run a real task in Cowork and watch it show up already knowing what to do.
What a time to be alive.
— Zach
One Last Thing
And it might be the best part
This book is the map. But the world we just talked about, the one where the tools change every single week? You can't keep up with that from a guide. You keep up with it from a person who does the keeping-up for you.
So here's the good news: by grabbing this book, you're now subscribed to The Filter, my free weekly newsletter. That wasn't a trick. It's genuinely the most valuable thing I give away, and I'd feel bad not getting it into your hands.
Here's what it actually is. Every Friday morning, I sit down for four or five hours and write the whole thing out by hand. No AI writing it, no team, no recycled press releases. Just me, sifting through everything that happened in AI that week, testing the tools in real time, and writing out my honest thoughts on what's worth your attention and what's noise. Readers tell me it's one of the most valuable AI things they read all week, and that's the bar I'm chasing every single Friday.
That's how you stay ahead. The book gives you the foundation. The Filter keeps it current, week after week, while everyone else drowns in the firehose.
Want a taste before Friday? You can read past editions and grab every other free guide I've made right here:
See all my free resources and past editions →
Scroll through, pick what's useful, and I'll see you in your inbox this Friday.
Talk soon,
— Zach
Work with Me
Want me on your problem for an hour?
Book a 1:1 call and we build the fix live — the workflow you want automated, the tool you can’t crack, whatever’s stuck. Direct, hands-on, no pitch waiting at the end.